The bacterial enzyme L-asparaginase (L-ASP) was approved by the FDA in 1979 for treatment of acute lymphoblastic leukemia, and it is still being used in standard clinical practice for that indication. In the early 1990s a pegylated version of L-ASP proved more favorable in terms of pharmacokinetic and immunological properties. Our studies have provided a rationale for possible use of the enzyme to treat a subset of ovarian cancers and possibly other solid tumor types. Until there are clinical data, that remains simply an unproven hypothesis, but the clinical possibility is directly in line with the drive toward biomarkers and personalization of medicine. The following are our research steps in the project: Microarray molecular profiling studies - In collaboration with the Brown/Botstein group at Stanford, we used cDNA microarrays to profile RNA expression in the 60 human cancer cell lines (the NCI-60) used by the NCIs Developmental Therapeutics Program to screen >100,000 compounds for anticancer activity since 1990. Analysis of the data showed a very strong, highly statistically significant negative correlation (r = -0.98) between the activity of L-ASP and the expression of asparagine synthetase (ASNS). That is, leukemia cell types expressing low ASNS were more sensitive to L-ASP. That result made sense in terms of the existing concept that L-ASP works by selectively starving the leukemia cells of asparagine. But we also found, surprisingly, that there was a negative correlation (r = -0.89) in the case of the ovarian cell lines. We later corroborated that negative correlation using three other microarray platforms: Affymetrix HU-6800 with the Lander/Golub group at what is now the Broad Institute (r = 0.90), Affymetrix HG-U95 with Scherf/Dolginow at GeneLogic (r = -0.85), and Affymetrix HG-U133 with Kaldjian/Scherf at GeneLogic (r = -0.86). In addition, array CGH with Joe Grays group at UCSF showed a highly negative correlation between L-ASP activity and DNA copy number in the chromosomal region of ASNS (r = -0.98). That observation reinforced the gene expression findings and also suggested that copy number differences were instrumental in determining expression of ASNS. siRNA studies for causality. The microarray studies indicated a relationship between L-ASP and ASNS, but they couldnt show whether the relationship is causal or epiphenomenal. To address that question, postdoctoral fellow Philip Lorenzi collaborated with Natasha Caplens group (GSS, OSTP, OD, CCR, NCI) to establish a pharmacological assay system with siRNA knock-down of ASNS. Two different siRNAs consistently reduced basal expression of ASNS about five-fold at the RNA level (by branched-DNA assay) and protein level (by quantitative Western blot) in all cell lines tested. In initial pharmacological tests, ovarian cell lines OVCAR-3 and OVCAR-4 showed potentiation of L-ASP activity about proportional to the knock-down (i.e., about five-fold). That result was gratifying, but the surprise was OVCAR-8, the ovarian line lowest in basal ASNS expression: Five-fold knock-down of ASNS produced an 600-fold potentiation of L-ASP activity. The next question was whether classical multi-drug resistance, the major barrier to long-term effectiveness of anticancer agents such as paclitaxel, would extend to L-ASP. To address that question, we first asked whether OVCAR-8/ADR-RES, a derivative of OVCAR-8 selected for classical multi-drug resistance, would show upregulated ASNS or resistance to L-ASP. The answer was no. Furthermore, siRNA knock-down of ASNS in the resistant line showed the same 600-fold potentiation of L-ASP activity as had the parental OVCAR-8. Immunohistochemistry (IHC): In collaboration with Paul Goldsmith (APPU, OSTP, OD, CCR, NCI), we used ASNS peptides to immunize rabbits and develop the first effective, selective antibodies against ASNS. (In a CCR collaboration with Becton-Dickinson led by Shoshana Segal and David Goldstein (OSTP, OD, CCR, NCI), we also developed mouse monoclonal anti-ASNS antibodies, but the rabbit antibodies are the ones weve used for the studies to date.) In a collaboration with Mark Raffeld (LP, CCR, NCI), we fixed and embedded pellets of the ovarian cell lines for IHC. The strength of staining was in the order expected on the basis of the microarray data, and, significantly, essentially all cells in any given culture stained essentially equally, independent of cell cycle or other physiological differences. Tissue arrays: To project the studies in cell lines toward the clinic, we then collaborated with Mark Raffeld using tissue arrays composed of hundreds of different ovarian and other cancer types. Approximately 15% of the clinical ovarian cancers showed low staining for ASNS (0 to 1+). Of course, absent clinical response data, we have no functional measure of how low the ASNS level has to be for clinical purposes. Importantly from the perspective of therapy, staining was essentially uniform across all tumors cells in each specimen; if some tumor cells in a specimen were low-expressers, all were low. Clinical trials: An advantage of finding new uses for old drugs is that the path to clinical testing is much shorter than it is for new agents. Currently in progress at three institutions (PI: Dan Von Hoff at TGen) is a Phase I study of pegylated L-ASP (OncosparTM) in combination with gemcitabine. Ivan Horak (Enzon, Inc.) and I are co-investigators in that trial and the others to be mentioned below. Enzon is the supplier of Oncospar. That trial is providing background information to complement the large body of pharmacokinetic, toxicological, and oncological information already in the literature on use of Oncospar to treat leukemias. Two Phase II single-agent studies of Oncospar are in the planning stages, one an all-comers trial with Dan Von Hoff as PI and one ovarian cancer trial at the NIH Clinical Center with Elise Kohn (MOB, CCR, NCI) as PI. Other studies in progress: Additional ovarian cancer cell lines - To test the L-ASP/ASNS relationship in larger numbers of ovarian cell lines, we are collaborating with Michael Birrer on 25 lines that he has collected. ASNS expression is being assessed in those lines at the RNA level by Affymetrix HG-U133 microarray and branched-DNA assay and at the protein level by a liquid-phase sandwich assay developed in collaboration with Paul Goldsmith as well as by IHC in collaboration with Mark Raffeld. From a clinical perspective, the IHC is most pertinent since that will be the principal clinical pathology test to be used later on in selecting patients. Cell lines other than ovarian We are extending the siRNA and staining studies to other cell line types, including breast, lung, and melanoma. Bioinformatic pathway analysis We've analyzed pathways related to ASNS to identify other molecules that may affect L-ASP activity and may point the way to effective combinations of L-ASP with other drugs. Shotgun siRNA studies Toward the same ends, weve compiled (and ordered) a library of several hundred molecular targets for shotgun siRNA studies and have established (in collaboration with Natasha Caplen) a fluent pharmacological assay system (in 348-well format) to test the effect of those knock-downs on L-ASP activity. Metabolomic studies Toward those same ends, we are collaborating with Felice de Jong at Metabolon, Inc. to [summary truncated at 7800 characters]